AI doesn’t threaten universities because it can explain concepts. It threatens them because higher education already turned itself into a content pipeline. If learning is just information transfer rather than formation in real social contexts, AI simply finishes what the system started. What then breaks is the link between how knowledge is produced and how people actually develop.
Agreed - it's a huge change for the Unis, given they are concerned themselves with knowledge, and AI is in part about automating knowledge. Some will use the change to adapt, improve and thrive, others will not. (and die; or die trying) Have you seen this 'the retvrn of the oral exam', only this time *scalable* - courtesy of the (hated) AI? (apology for the noise if the NYT article referencing 'oral exam' mentions it - can't see its content.)
I love how it turns out the solution to cheating on written exams to be - bringing back the oral Q&A exam of old. That was 'the gold standard' when I attended uni. (decades ago)
If only the spineless, vapid and stupid universe administrators could be replaced by Opus-5/GPT-6/Gemini-4, now that would be win^3. (money $$$, malign influence, efficiency)
Aside - remembered this David Brooks 'Mea Culpa' ("I have a confession to make - I'm a member of the educated elites...") talk -
Love the idea of using AI to fight against AI cheating or cognitive outsourcing! Jimmy Alfanso Licon also gives some examples of that. The tricky part is when it ends up being a lot of work on profs to implement. Learning curves are one hell of a deterrent.
Lots of work for Profs to implement - yes true, but only this time, the 1st time, guerrilla-time going 0->1. I expect AI Cos to offer a ready made turnkey solutions. I mean - if a single lone Prof could wire this up for $20 total, have it run, and run well - imagine what Anthropic or OpenAI or Google can offer? As a turnkey solution offered to the Unis for real $$$s? The mind boggles.
We all find it hard to imagine and esp *internalise* it. Maybe easier if we are to compare-and-contrast with what's already happened in s/w development in the 3 short months Oct-Dec 2025. And people became aware during the Xmas break, when they had some down time on their hands to try the coding agents themselves fr. Myself I became aware of the change in Sep'25 - had a coding challenge in a language I dislike (and not respect) with crucially "any and all AI help allowed" (explicitly; to this firm's great credit). Now I write zero code myself, and read <50% of the code Codex and Claude write for me on my instructions. I expect the code I read to be ~0% next. (same way I don't read the assembly code the compiler produces)
A lot I agree with here, but if the alternative to going to college is getting a job or playing more baseball, what's the net value added? Are profs better at formation than employers or coaches?
Depends on the prof, the job, and the coach. I’m not someone who thinks college is for everyone. For some people, a good job or sport (or military) is a better path right out of high school. But, to be clear, I don’t think the professors are the only ones doing formation. Just as the whole sport forms the player, the whole university experience is transformative. It’s not for everyone, but I believe it’s up there with a good team, start-up, or battalion.
So students learn to avoid "groupthink" by practicing thinking with others and communing with peers and mentors in groups? I think you put your finger on one of the university's most basic contradictions.
I don’t see groupthink as thinking in groups; it’s uncritically accepting what like-minded people believe. College provides an opportunity for students to build critical thinking skills and relationships with fellow students and professors who aren’t like-minded. Some campuses might lack viewpoint diversity, but that’s a separate problem. The best university environments aren’t echo chambers.
great essay (and great audio!)
AI doesn’t threaten universities because it can explain concepts. It threatens them because higher education already turned itself into a content pipeline. If learning is just information transfer rather than formation in real social contexts, AI simply finishes what the system started. What then breaks is the link between how knowledge is produced and how people actually develop.
Agreed - it's a huge change for the Unis, given they are concerned themselves with knowledge, and AI is in part about automating knowledge. Some will use the change to adapt, improve and thrive, others will not. (and die; or die trying) Have you seen this 'the retvrn of the oral exam', only this time *scalable* - courtesy of the (hated) AI? (apology for the noise if the NYT article referencing 'oral exam' mentions it - can't see its content.)
https://www.behind-the-enemy-lines.com/2025/12/fighting-fire-with-fire-scalable-oral.html
I love how it turns out the solution to cheating on written exams to be - bringing back the oral Q&A exam of old. That was 'the gold standard' when I attended uni. (decades ago)
If only the spineless, vapid and stupid universe administrators could be replaced by Opus-5/GPT-6/Gemini-4, now that would be win^3. (money $$$, malign influence, efficiency)
Aside - remembered this David Brooks 'Mea Culpa' ("I have a confession to make - I'm a member of the educated elites...") talk -
https://www.youtube.com/watch?v=QSa52TR9tCA
Worth sharing I think.
Love the idea of using AI to fight against AI cheating or cognitive outsourcing! Jimmy Alfanso Licon also gives some examples of that. The tricky part is when it ends up being a lot of work on profs to implement. Learning curves are one hell of a deterrent.
https://open.substack.com/pub/jimmyalfonsolicon/p/synthetic-socrates-teaching-assistant
Good link - thanks. Yeah - that's the dream, right? Socratic dialogue. Not the 1st ones to desire that us. :-) "Steve Jobs on Al in 1985" comes to mind https://www.youtube.com/watch?v=A2Au922DQ84. (same idea, maybe the same anecdote too https://www.youtube.com/shorts/By566GHmA7g)
Lots of work for Profs to implement - yes true, but only this time, the 1st time, guerrilla-time going 0->1. I expect AI Cos to offer a ready made turnkey solutions. I mean - if a single lone Prof could wire this up for $20 total, have it run, and run well - imagine what Anthropic or OpenAI or Google can offer? As a turnkey solution offered to the Unis for real $$$s? The mind boggles.
We all find it hard to imagine and esp *internalise* it. Maybe easier if we are to compare-and-contrast with what's already happened in s/w development in the 3 short months Oct-Dec 2025. And people became aware during the Xmas break, when they had some down time on their hands to try the coding agents themselves fr. Myself I became aware of the change in Sep'25 - had a coding challenge in a language I dislike (and not respect) with crucially "any and all AI help allowed" (explicitly; to this firm's great credit). Now I write zero code myself, and read <50% of the code Codex and Claude write for me on my instructions. I expect the code I read to be ~0% next. (same way I don't read the assembly code the compiler produces)
A lot I agree with here, but if the alternative to going to college is getting a job or playing more baseball, what's the net value added? Are profs better at formation than employers or coaches?
Depends on the prof, the job, and the coach. I’m not someone who thinks college is for everyone. For some people, a good job or sport (or military) is a better path right out of high school. But, to be clear, I don’t think the professors are the only ones doing formation. Just as the whole sport forms the player, the whole university experience is transformative. It’s not for everyone, but I believe it’s up there with a good team, start-up, or battalion.
So students learn to avoid "groupthink" by practicing thinking with others and communing with peers and mentors in groups? I think you put your finger on one of the university's most basic contradictions.
I don’t see groupthink as thinking in groups; it’s uncritically accepting what like-minded people believe. College provides an opportunity for students to build critical thinking skills and relationships with fellow students and professors who aren’t like-minded. Some campuses might lack viewpoint diversity, but that’s a separate problem. The best university environments aren’t echo chambers.
They're necessarily echo chambers. Humans can't think in groups without uncritically accepting what like-minded people believe..
Bold claim! If true, our deeply social species has bigger problems—except for the few hermits among us.